Machines Will Be Smarter Than Humans Within 30 Years, SingularityNet’s Founder Says

Ben Goertzel wants to make sure AI doesn’t wipe out humanity

'Benevolent superhuman AGI can create utopia for humans and machines alike,' Goertzel says.

Ben Goertzel gives us roughly five to 30 years before machines are smarter than humans. And he’d know. When Goertzel says he is creating benevolent, decentralized artificial general intelligence (AGI), he means he’s working on machines that will be more rational, compassionate and broad-minded than their human counterparts in order to eventually help regulate society—and potentially even solve some of the problems we haven’t been able to fix ourselves.

“Benevolent superhuman AGI can create utopia for humans and machines alike,” Goertzel, founder and CEO of blockchain-based AI marketplace SingularityNet, said. “Nasty AGI at the human level, or even slightly infrahuman level, could create dystopia or even wipe out humanity. This is about as important as things get—within the human corner of reality, at any rate.”

Sophia the Robot
Getty Images

Since 2015, Goertzel has been the chief scientist of Hanson Robotics, a company perhaps best known for Sophia the Robot. Thanks to a combination of advanced technologies, this human-like robot can recognize faces and expressions, provide unique responses to situations and even express feelings. Goertzel said the next commercial steps for Sophia will be mass production, followed by a rollout as a service robot for offices, stores, schools, hospitals and eventually homes. “To be an effective service robot, she needs not only a mass-produced body, but also increased general intelligence,” he said. “We’re working on it.”

Goertzel’s career began in academia. After earning his Ph.D. in mathematics, he became a math professor at the University of Nevada, Las Vegas before teaching computer science at the University of Waikato in New Zealand, cognitive science at the University of Western Australia and computer science at the College of Staten Island. Since leaving academics in 1997, he’s been mostly focused on entrepreneurial AI projects.

That includes his current role at SingularityNet, which is working on an app store for algorithms. That means developers will be able to monetize AI services without organizations like Google or Amazon, and buyers of said services won’t be dependent on agencies, which can be expensive. Available services include image and emotion recognition, object detection and speech recognition, with more to come in ensuing months.

“I, for one, look forward eagerly to humanity being subtly guided along by compassionate genius machines,” he said.

Big Mistake

In 1997, Goertzel recounted, “I was planning to launch an AI-driven consumer web search engine, but was talked out of it by a series of New York VCs who told me there would never be any money to be made in consumer web search.”

Lesson Learned

“I learned that much of the time, experts in the current business world are not really experts in the business world as it’s going to be during the next few years,” he said.

Job Profile

In the next 90 days, Goertzel is focused on a software spinoff that will bring corporations onto the SingularityNet platform. Over the longer term, he wants to focus on scientific and technical aspects of creating AGI. “For most of my AI career, AI was an obscure and unpopular pursuit. But in the last few years, things have shifted,” he explained. “Now, my skills, knowledge and insights are in extraordinarily high demand, which is actually a bit disorienting.”

How He Got the Gig

Goertzel founded SingularityNet in 2017 after a nearly three-decade career in education and entrepreneurship. “The part of me that wants to build stuff and directly impact the world really got frustrated with the slow pace of academia and the lack of opportunity there to work tightly with teams of trained engineering professionals expert at implementing ideas in reality,” he said.

This story first appeared in the March 18, 2019, issue of Adweek magazine. Click here to subscribe.

Recommended articles